Treatment effect estimation for time-to-event data with intercurrent events
Date:
Analyzing treatment effects in clinical trials with time-to-event outcomes is complicated by the presence of intercurrent events. In randomized controlled trials and observational studies that focus on time-to-event endpoints, intercurrent events can arise in two ways: as semi-competing events, which modify the hazard of the primary outcome events, or as competing events, which make the definition of the primary outcome events unclear. In this work, we will show how to define, estimate, and make inferences concerning objectives with causal interpretations. We derive the mathematical formulations of the causal estimands corresponding to the five strategies outlined in ICH E9 (R1) and clarify the data structure required to identify them. Furthermore, we introduce nonparametric and semiparametrically efficient methods for estimating and making inferences about these causal estimands, including the asymptotic variance of estimators and the construction of hypothesis tests. We develop an R package “tteICE” with a Shiny interface.